Abstract
Background
Methods
Results
Conclusions
Introduction
The host immune response to HBV infection plays a pivotal role in determining whether acute infection resolves or develops into chronic hepatitis B (CHB) infection. Individuals who spontaneously clear HBV infection following an acute infection display a vigorous, polyclonal, HBV-specific CD8+ and CD4+ T-cell response [1]. In contrast, the persistent viral antigenic stimulation in CHB patients results in dysfunctional HBV-specific T-cell and B-cell responses and impaired natural killer (NK) cell antiviral function [2].
Toll-like receptors (TLRs) are a family of membrane-bound pattern receptor recognition receptors that mediate both innate and adaptive immune responses by recognizing pathogen-associated molecular patterns. TLR8 is located in the endosomal membrane of immune cells, including monocytes, macrophages and dendritic cells (DCs), and recognizes single-stranded RNA [3,4]. Activation of TLR8 leads to secretion of inflammatory cytokines, chemokines and interferons (IFNs), and initiates innate and adaptive immune responses [5–11], making TLR8 an attractive target for drug development for the treatment of CHB [12,13].
Selgantolimod (GS-9688) is a small molecule oral TLR8 agonist in clinical development for the treatment of CHB. Selgantolimod is selective for TLR8, induces production of immune mediators such as interleukin (IL)-12 and IL-18, antiviral cytokines tumour necrosis factor (TNF)-α and IFN-γ, and various proinflamma-tory and anti-inflammatory cytokines (IL-1α, IL-1β, IL-6 and IL-1RA) in human peripheral blood mono-nuclear cells with nanomolar potency, while exhibiting minimal effects on the TLR7-induced cytokine IFN-α [14]. In a woodchuck model of CHB, oral administration of 3 mg/kg selgantolimod once weekly for 8 weeks led to a reduction of serum woodchuck hepatitis virus (WHV) DNA and woodchuck hepatitis surface antigen (WHsAg) in four of six animals. This response was sustained in three animals and resulted in detectable anti-WHsAg antibodies and enhanced peripheral WHV-specific T-cell responses [15]. While preclinical results are promising, in-human studies are needed to validate this target in HBV.
Preclinical PK studies of selgantolimod demonstrated it has high absorption, good solubility and high forward permeability with no efflux transport. Selgantolimod had low bioavailability in nonclinical species, consistent with high first-pass metabolism. Studies in portal vein cannulated dogs and monkeys demonstrated high absorption and high hepatic extraction. Metabolism is the major route of elimination of selgantolimod as determined by preclinical in vivo mass balance and in vitro metabolism studies.
This study was a first-in-human, randomized, blinded, placebo-controlled, single ascending dose study of sel-gantolimod in healthy volunteers. We present the safety, tolerability, pharmacokinetics (PK) and pharmacodynamic (PD) effects on various cytokines, chemokines and acute phase proteins of selgantolimod administered as a single oral dose of 0.5, 1.5, 3 or 5 mg, as well as PK/PD analyses.
Methods
Study design
This Phase I, randomized, blinded, placebo-controlled study was conducted December 2016 through March 2017 at one site in Auckland, New Zealand (Australian New Zealand Clinical Trials Registry, ACTRN12616001646437) to evaluate the safety, tolerability, PK and PD of single ascending oral doses of selgantolimod. The study consisted of two parts: one evaluating single ascending doses and one evaluating the effect of food on PK and PD. In each of the four cohorts evaluating single ascending doses, 15 subjects were randomized 4:1 to receive selgantolimod (0.5, 1.5, 3 or 5 mg) or matching placebo in a fasted state. For evaluation of food effect, an additional cohort of 12 subjects received a single dose of 1.5 mg selgantolimod in a fasted sate, followed by a 6-day washout, and a single dose of 1.5 mg selgantolimod administered with a moderate-fat meal (approximately 600 kcal, 25% to 30% fat).
Subjects
Eligible subjects were healthy volunteers, 18 to 45 years of age, with no known liver disease. Subjects had a body mass index of 19 to 30 kg/m2, creatinine clearance ≥90 ml/min, and either normal 12-lead electrocardiograms (ECGs) or ones with abnormalities considered clinically insignificant by the investigator. Exclusion criteria included: HBV, HCV or HIV infection; use of systemic steroids, immunosuppressant therapies or chemotherapeutic agents within 3 months of study; and history of significant medical condition. Concomitant prescription or nonprescription medications, with the exception of vitamins, acetaminophen and/or ibuprofen, were prohibited during the study unless approved by the study team.
All subjects provided written informed consent before undertaking any study-related procedures. The protocol was approved by the local ethics committee, and the study was conducted in accordance with Good Clinical Practice and the principles of the Declaration of Helsinki.
Safety assessments
Safety was assessed by adverse events (AEs) and concomitant medications, clinical laboratory analyses, vital signs measurements, ECGs, physical examinations and ophthalmic examinations.
AEs and concomitant medications were monitored from screening until follow-up (14 ±2 days after last dosing of study drug). Other safety assessments were conducted at screening; days -1, 1, 2, 4, 5, 7 (and days 8, 9, 10 and 14 for the food effect cohort); and the follow-up visit (10 ±1 days and 14 ±2 days after last dosing of study drug for the single ascending doses and food effect cohorts, respectively).
Pharmacokinetic and pharmacodynamic assessments
Plasma samples were collected within 5 min before dosing and at 0.25, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 8, 12 and 24 h after dosing for PK assessments. For PD analysis, in the single ascending dose cohorts, serum samples were collected predose and at 1, 2, 4, 8, 12 and 24 h after dosing (matched with plasma PK sample collection time points); a serum sample was also collected at 144 h after dosing. In the food effect cohort, serum samples were collected predose and at 1, 2, 4, 8, 12 and 24 h after dosing in the fasted and fed states (matched with plasma PK sample collection time points); a serum sample was also collected at 144 h after dosing in the fed state.
Analysis of PK and PD samples
Analysis of PK plasma samples was performed by an independent contract research facility (Frontage Laboratories, Inc., Exton, PA, USA) using a validated liquid chromatography tandem mass spectroscopy method. Selgantolimod and an isotopically-labelled internal standard (GS-829520) in plasma were extracted with acetonitrile, separated on a C18 reversed phase column, eluted with mobile phases consisting of 10 mM ammonium acetate in water and acetonitrile at a flow rate of 0.6 ml/min, ionized by positive ion electrospray and detected in the multi-reaction monitoring mode using precursor → product ions of m/z 293.9 → 180.2 and m/z 302.9 → 180.2 for selgantolimod and an isotopically labelled internal standard (GS-829520), respectively. The bioanalytical method was validated over a calibrated range of 0.5-500 pg/ml. Interassay precision, based on coefficient of variation (%CV), ranged from 2.8% to 9.2%, and accuracy ranged from -5.3% to -1.3% based on interday percent relative error. All samples were analysed in the timeframe supported by frozen stability storage data.
The PD effects of selgantolimod were evaluated by analysis of serum levels of cytokines (IL-12p40, IL-12p70, IL-1RA, IFN-γ and TNF-α), chemokines (C-C motif chemokine ligand [CCL] 2 [MCP-1], CCL4 [MIP-1p], CCL8 [MCP-2], CCL11 [eotaxin-1], CCL20 [MIP-3a], C-X-C motif chemokine ligand [CXCL] 8 [IL-8], CXCL9 [MIG] and CXCL10 [IP-10]) and acute phase proteins (C-reactive protein [CRP] and serum amyloid A [SAA]). Serum concentrations of IL-12p40 and IL-1RA were measured by high-sensitivity enzyme-linked immunosorbent assays (R&D Systems, Minneapolis, MN, USA). Serum concentrations of all other cytokines, chemokines and acute phase proteins were measured using multiplexed Luminex assays (Myriad RBM Inc., Austin, TX, USA), except for IL-12p70 and IFN-γ, which were measured using high-sensitivity Cira-plex assays (Aushon Biosystems, Billerica, MA, USA).
Statistical analyses
The sample size for this study was determined based on past experience with similar types of studies that would provide a suitable assessment of the safety, PK and PD profiles of selgantolimod. No formal sample size calculation was performed.
AEs, laboratory test results, and treatment-emergent laboratory abnormalities were summarized by treatment group for the single ascending dose cohorts and by treatment period for the food effect cohort.
Selgantolimod PK parameters were estimated using the linear/log trapezoidal rule with standard noncom-partmental methods and summarized using descriptive statistics by treatment. Dose proportionality of primary PK parameters (area under the plasma concentration time curve [AUC] from time zero to the last quantifiable concentration [AUClast], AUC from time zero to infinity [AUCinf] and maximum concentration [Cmax]) was assessed using a power model. The power model was based on regression of natural log-transformed PK parameters on natural log-transformed dose (y = β0 x doseβ1) and used to calculate the population mean slope β1 and corresponding 90% confidence intervals (CIs). For completeness, dose proportionality was also assessed using an analysis of variance (ANOVA) model to calculate the ratio of geometric least-squares means (GLSMs) and corresponding 90% CIs for each primary PK parameter, with dose as a fixed effect and PK parameters dose-normalized to a reference dose of 1.5 mg (results not presented). The effect of food on sel-gantolimod PK was assessed using an ANOVA model to calculate the GLSM ratio (fed/fasted) and corresponding 90% CIs for AUClast, AUCinf and Cmax, with treatment as a fixed effect.
PD analyses of biomarkers (cytokines, chemokines and acute phase proteins) included determination of predose and postdose values, change from predose and fold-change from predose. Values were summarized by treatment group for the single ascending dose cohorts and by treatment period for the food effect cohort.
The PK/PD relationship between selgantolimod exposure and changes in biomarker response was assessed by plotting the log-transformed ratio of biomarker AUC from time zero to 24 h (AUC0–24) for after selgantolimod treatment and at baseline (log10[AUC0–24, treatment/AUC0–24, baseline]) versus selgantolimod AUC0–24. A 3-parameter maximum effect (Emax) model (y = E0 + Emax x x / [EC50 + x]), with E0 fixed at log10(24) (that is, no fold change in biomarker AUC0–24), was used to determine Emax and the half-maximal effective concentration (EC50) of selgantolimod.
Results
Disposition and baseline characteristics
A total of 71 subjects were randomized into the study, including 59 in the single ascending dose cohorts (48 selgantolimod subjects [12 in each of 4 cohorts] and 11 placebo subjects) and 12 subjects in the food effect cohort. All subjects completed the study. Baseline characteristics were generally similar across treatment groups. Overall, the majority were male (65%) and White (78%), with a median age of 24 years (Table 1).
Demographics and baseline characteristics
Continuous data are presented as median (Q1, Q3). BMI, body mass index.
Safety
Selgantolimod was generally safe at all doses evaluated. No SAEs or AEs leading to study drug discontinuation were reported. In the single ascending dose cohorts, treatment-emergent AEs were reported in 29 of 48 selgantolimod subjects (60%) and 5 of 11 placebo subjects (46%; Additional file 1). There was a dose-dependent trend in AEs, with the highest incidence in the 5 mg group (12 of 12 subjects [100%]). All AEs were grade 1 in severity, with the exception of two grade 2 AEs (diarrhoea and arthropod bite), both of which resolved. Treatment-related AEs were reported in 24 of 48 selgantolimod subjects (50%) and 2 of 11 placebo subjects (18%). The most common treatment-related AEs in selgantolimod subjects were nausea (38%); vomiting (25%); and abdominal discomfort, diarrhoea and headache (each 10%). The most common treatment-related AE in placebo subjects was nausea (18%). Among the 12 subjects in the food effect cohort, treatment-emergent AEs were reported in 8 subjects (67%) and 4 subjects (33%) in the fasted and fed states, respectively. Treatment-related AEs were reported in six subjects (50%), all in the fasted state. The only treatment-related AE reported in ≥2 subjects in the food effect cohort was nausea (four subjects [33%]).
In the single ascending dose cohorts, 33 of 48 selgantolimod subjects (69%) and 5 of 11 placebo subjects (46%) experienced laboratory abnormalities, the majority of which were grade 1 in severity (Additional file 1). No grade 3 or 4 laboratory abnormalities were reported. Grade 2 laboratory abnormalities in selgantolimod subjects included elevations in alanine aminotransferase (ALT; two subjects) and low-density lipoprotein, lipase and total cholesterol (each one subject).
The grade 2 elevations in ALT occurred in two subjects in the 5 mg cohort. Both subjects had normal ALT at baseline and grade 2 elevations on day 2, which decreased over the course of the study to within normal range at the follow-up visit. Both subjects also had concurrent grade 1 elevations in aspartate aminotransferase (AST) and/or total bilirubin, with otherwise normal values for these analytes. An additional subject in the 5 mg cohort with normal ALT at baseline had grade 1 elevated ALT at all study visits and at follow-up; the subject's ALT was within normal range 1 week after follow-up. No graded ALT elevations occurred in the other single ascending dose cohorts.
Among the 12 subjects in the food effect cohort, graded laboratory abnormalities occurred in 5 subjects (42%) and 3 subjects (25%) in the fasted and fed states, respectively, all of which were grade 1.
No clinically relevant changes in vital signs or clinically significant ECG or ophthalmic abnormalities were observed during the study.
Based on these data, the 1.5-mg and 3-mg doses of GS-9668 have been selected for evaluation in future clinical studies.
Pharmacokinetics
In the single ascending dose cohorts, selgantolimod plasma concentrations peaked at approximately 0.5 to 1 h (median time taken to reach maximum concentration [Tmax]) and then decreased gradually over 24 h (Figure 1A). Mean selgantolimod exposure increased with increasing dose (Additional file 2). Median terminal half-life ranged from 4.3 to 11.3 h. Across the evaluated dose range, increases in selgantolimod exposures were approximately dose proportional using power model analysis (mean slope [90% CI]: 0.97 [0.81,1.13], 0.87 [0.71,1.04] and 0.98 [0.79,1.16] for AUClast, AUCinf and Cmax, respectively).
Pharmacokinetics of selgantolimod
Administration of selgantolimod with a moderate-fat meal led to a delay in absorption; median fed-state Tmax was 1.5 h compared with a fasted-state Tmax of 0.5 h (Figure 1B and Additional file 2). Mean selgantolimod exposure after dosing with a moderate-fat meal was slightly higher compared with the fasted state, with increases in AUClast, AUCinf and Cmax of 23%, 18% and 8%, respectively.
Pharmacodynamics
To assess PD response to selgantolimod administration, an initial focus was placed on measurement of serum levels of IL-12p40 and IL-1RA, as these two cytokines have been identified to be strongly induced in non-human primate (NHP) studies following oral sel-gantolimod treatment. Mean serum concentrations of IL-12p40 and IL-1RA increased with increasing selgantolimod dose, reaching peaks approximately 4 h after dosing and decreasing to near predose levels at the 24 h time point (Figure 2A). Dose-dependent maximum induction was observed 4 h after dosing in the fasted state, with the greatest median fold change from baseline at the 5-mg dose (8-fold and 32-fold for IL-12p40 and IL-1RA, respectively; Table 2). Of note, PD responses were detected even at the lowest dose tested (0.5 mg). In contrast, subjects who were dosed with placebo had very modest fluctuations in levels of these PD biomarkers over time. A pronounced reduction in the magnitude of PD response was observed in subjects who were dosed with 1.5 mg selgantolimod in the fed state compared with the equivalent dose in the fasted state (Figure 2B). The maximum median fold change from baseline for IL-12p40 was 2.7-fold and 1.3-fold, and for IL-1RA was 2.5-fold and 1.3-fold, in the fasted and fed states, respectively (Table 2).
Dose-dependent induction of IL-12p40 and IL-1RA
Biomarker responses following single-dose administration of selgantolimod
The maximum fold change from baseline (median [min, max]) in serum concentration for each biomarker during the 24-h postdose time period is provided. CCL, C-C motif chemokine ligand; CRP, C-reactive protein; CXCL, C-X-C motif chemokine ligand; IFN, interferon; IL, interleukin; IP, interferon gamma-induced protein; MCP, monocyte chemotactic protein; MIG, monokine induced by gamma interferon; MIP, macrophage inflammatory protein; ND, not determined; SAA, serum amyloid A; TNF, tumour necrosis factor.
To further explore the breadth of the PD response, serum levels of a broader panel of cytokines, chemokines and acute phase proteins were measured. There was a clear dose-dependent increase in median fold change from baseline in serum concentrations of IL-12p70, IFN-γ, CCL2, CCL4, CCL8, CCL11, CCL20, CXCL8, CXCL9, CXCL10, CRP and SAA (Table 2). Notably, the kinetics of induction varied across this panel of biomarkers. The maximum fold change from baseline for most biomarkers, including IL-12p70, IFN-γ, CCL2, CCL8, CCL11, CCL20 and CXCL10, occurred at 4 h postdose. In contrast, TNF-α, CCL4 and CXCL8 reached maximum induction at 2 h postdose and CXCL9 peaked at 8 to 12 h postdose. The acute phase proteins CRP and SAA exhibited a substantially distinct profile, reaching maximum levels at 24 h postdose. For all assessed biomarkers, induction in the fed state was greatly diminished compared with the level of induction in the fasted state (Table 2).
Pharmacokinetics/pharmacodynamics relationship
The PK/PD relationship between selgantolimod exposure and biomarker induction was assessed by plotting bio-marker log10(AUC0–24, treatment/AUC0–24, baseline) versus selgantolimod AUC0–24. Biomarker responses increased with increasing selgantolimod exposure (Figure 3). With the exception of CCL2 and TNF-α, the exposure–response relationship for each biomarker demonstrated saturability that was adequately described by an Emax model. Based on the Emax models, near-saturation of IL-12p40, CCL4, CCL8 and CXCL9 induction occurred at a dose of approximately 5 mg, suggesting that increasing the dose beyond 5 mg is unlikely to result in further substantial serum increases in these biomarkers. In contrast, saturation of IL-1RA, CRP, CXCL10 and SAA appears to occur at doses greater than 5 mg. For CCL2 and TNF-α, the exposure–response relationship was linear within the evaluated dose range. Of note, the lower sel-gantolimod doses (0.5 and 1.5 mg) did not lead to a measurable increase in TNF-α levels (Figure 3).
Pharmacokinetic/pharmacodynamic relationship between fold change in biomarkers and selgantolimod exposure
Discussion
The host immune response to HBV infection plays a central role in whether acute infection is resolved or becomes chronic. CHB is characterized by dysfunction of both innate and adaptive immunity, and so activation of both arms of the immune response with a selective TLR8 agonist is a rational approach to reinvigorate host antiviral immunity. However, limiting systemic activation of this pathway is likely to be an important factor in achieving this aim. Therefore, sel-gantolimod was designed as an oral and selective TLR8 agonist with predicted high first pass clearance to limit systemic exposure. This study is the first report on the clinical safety, tolerability, PK and PD of selgantolimod.
Doses of 0.5 to 5 mg selgantolimod were safe and well tolerated up to 3 mg, with decreased tolerability at 5 mg. Most treatment-emergent AEs were grade 1 in severity. The most common treatment-emergent AEs assessed as related to selgantolimod were gastrointestinal events (nausea and vomiting). The majority of laboratory abnormalities were mild in severity.
Selgantolimod was rapidly absorbed in the fasted state or with a moderate-fat meal. High clearance rates support low bioavailability in humans. These data are in line with the high first-pass clearance observed in preclinical species, and the low bioavailability and high apparent systemic clearance observed in cynomolgus monkeys [14].
Consistent with observations in preclinical models, IL-12p40 and other cytokines and chemokines were induced after single-dose administration of selgantolimod in the fasted state in the single ascending dose cohorts of this study. Peak induction of IL-12p40 and IL-1RA occurred 4 h postdose, consistent with the kinetics of cytokine induction observed in preclinical studies in cynomolgus monkeys [14]. In cynomolgus monkeys, IL-12p40 induction was observed in specimens that received oral selgantolimod, but not in specimens that received a low-dose intravenous infusion of selgantolimod, which suggests that presystemic activation of TLR8 in gut-associated lymphoid tissues is an important driver of IL-12p40 induction. Given that elevated systemic levels of cytokines and chemokines were detected in this study, our interpretation is that these immune mediators likely derive from activated immune cells residing in the intestinal tract and/or liver.
The robust induction of chemokines, including CCL2, CCL4, CCL8, CCL11, CCL20, CXCL8, CXCL9 and CXCL10 indicates the potential for orally administered selgantolimod to induce trafficking of activated T-cells, NK cells and myeloid cells into the liver. In addition to the induction of factors likely to impact immune cell trafficking into tissues, we observed a clear increase in effector cytokines associated with antiviral immunity, including IL-12p70, IFN-γ and TNF-α. Furthermore, selgantolimod treatment elevated the levels of acute phase proteins CRP and SAA, consistent with activation of hepatocytes, which are considered to be a primary although not exclusive source of these inflammatory markers.
Induction of all cytokines and chemokines measured was substantially diminished when selgantolimod was administered in the fed state, although the mechanism by which this effect occurs remains unknown. In both the fasted and fed states, the magnitude of PD responses varied across subjects at each dose level, as indicated by the wide range of maximal induction (Table 2). The source of this variability is not known. Analysis of genetic polymorphism rs3764880 in TLR8, which has been previously linked with TLR8 signalling, did not reveal any strong or consistent association between genotype and magnitude of PD response (data not shown).
Overall, these data indicate that TLR8 activation leads to differential induction of biomarkers. Further analysis of how biomarker induction correlates with efficacy and safety is warranted.
Augmentation of innate and adaptive immune responses is a potential means of inducing antiviral immunity in CHB patients. The results of this first-in-human study demonstrate that selgantolimod is safe and well tolerated at doses up to 3 mg, is rapidly absorbed with high first-pass clearance, and can effectively induce cytokine and chemokine production downstream of TLR8 activation in humans. Immune cell activation and antiviral activity of selgantolimod is being investigated in ongoing Phase II clinical studies in CHB patients. Ongoing studies are evaluating the efficacy and tolerability of selgantolimod both as mono-therapy and in combination with other novel antiviral and immunomodulatory agents.
Footnotes
Acknowledgements
Medical writing support was provided by Pouya Javidpour of Gilead Sciences (Foster City, CA, USA). This work was supported by Gilead Sciences. Presented in part at The Liver Meeting (American Association for the Study of Liver Diseases), San Francisco, CA, USA, 9 to 13 November 2018. Abstract 390. Australian New Zealand Clinical Trials Registration: ACTRN12616001646437.
MR, JDL, AHL, AG, EPG, AJ, RLM, JL, SKT, SD, JW, PW, TL, SPF, AM and PG are employees of and/or hold stock in Gilead Sciences. SK has received research grants from Merck, Gilead and Arbutus, and has participated on advisory boards for Merck. BP has received research funding from Gilead, paid to the University of Maryland. EJG has served as an advisory board member for AbbVie, Gilead, Janssen, Mylan, Roche and VIR Bio, and has participated in speakers’ bureaus for AbbVie and Gilead. NA reports no conflicts of interest.
